Layered firepower planning method fusing diffusion model and chaotic polynomial

Abstract

The invention discloses a layered firepower planning method fusing a diffusion model and a chaos polynomial, and belongs to the field of spacecraft manufacturing and application. The target damage expectation evaluation method including energy field rasterization unified matrix representation, diffusion model energy field prediction and chaotic polynomial damage expectation calculation is established, and the efficiency and convenience of damage effect evaluation are improved. And counting energy field information contained in each grid based on a rasterization preprocessing method of the maximum range of the energy field. A hierarchical optimized firepower planning architecture is established, aiming point planning efficiency is improved, a top layer determines a distribution relation matrix of the aircraft and the target for firepower distribution, meanwhile, the incidental damage effect of the aircraft on the target is considered by utilizing a grouping strategy, and on the basis, a bottom layer aims at each group; and aiming point optimization is carried out by using a rapid and accurate target damage expectation evaluation method, an aiming point planning result is rapidly obtained, and the aircraft efficiency is exerted in high real-time performance according to the aiming point planning result.

Description

Technical Field

[0001] This invention belongs to the field of spacecraft manufacturing and application, and specifically relates to a hierarchical optimization method of fusion diffusion model and chaotic polynomial. Background Technology

[0002] Aiming point optimization, which involves rationally allocating firepower resources and determining the optimal aiming point coordinates based on constraints such as target value, aircraft performance, and target threat, combined with real-time situational information, is one of the core issues in collaborative combat command and decision-making. Aiming point optimization is a nonlinear programming problem with multiple parameters and constraints. As the types and numbers of aircraft and targets increase, the number of possible solutions grows exponentially. Improving the solution speed and accuracy of aiming point optimization has become a major challenge that urgently needs to be addressed.

[0003] ① Traditional aiming point optimization methods often simplify the damage effect into a fixed value within a given damage range, and most of them do not consider the impact of uncertainties in the aircraft circular error probability (CEP) and target position error (TEP), resulting in large errors in the final aiming point optimization results and affecting the final combat effectiveness. However, considering uncertainties makes damage effect assessment very complex and time-consuming, seriously affecting the solution speed of fire planning and making it difficult to meet real-time requirements.

[0004] ② Existing methods do not consider the damage effect of aircraft on nearby targets, which will lead to a certain degree of waste of firepower resources, especially in non-saturated mission scenarios where firepower resources are limited.

[0005] ③ For unsaturated mission scenarios with insufficient aircraft, existing methods are not intelligent enough in the selection and allocation of mission targets, which may result in the damage effect of targets with high damage determination not meeting expectations. Summary of the Invention

[0006] The purpose of this invention is to provide a hierarchical fire planning method that integrates diffusion models and chaotic polynomials. This method fully considers the influence of the aircraft's CEP, the target's TEP, and collateral damage effects, and performs rapid calculation of the optimal aiming point for aircraft fire planning. This makes the fire planning results closer to real mission scenarios and improves the real-time performance of aircraft fire planning and aircraft effectiveness.

[0007] The objective of this invention is achieved through the following technical solution:

[0008] This invention discloses a hierarchical fire planning method that integrates diffusion models and chaotic polynomials. It establishes a target damage expectation assessment method that includes a unified matrix representation of the energy field rasterization, energy field prediction using a diffusion model, and damage expectation calculation using chaotic polynomials, thereby improving the efficiency and convenience of damage effect assessment. A hierarchical optimized fire planning architecture is established to improve the efficiency of aiming point planning. The top layer determines the allocation relationship matrix between the aircraft and the target for fire allocation, while simultaneously utilizing a grouping strategy to consider the collateral damage effects of the aircraft on the target. Based on this, the bottom layer optimizes the aiming point for each group using a fast and accurate target damage expectation assessment method, quickly obtaining the aiming point planning results. Based on the aiming point planning results, the aircraft's effectiveness can be maximized with high real-time performance.

[0009] This invention discloses a hierarchical firepower planning method based on a fusion diffusion model and chaotic polynomials, comprising the following steps:

[0010] Step one involves dividing the energy field into several square grids according to the length and width of the damage range of the aircraft's energy field, with predetermined side lengths. The energy field information contained within each grid is then statistically analyzed. For fragment energy fields, the number of fragments meeting the kinetic energy threshold within each grid is counted, resulting in a rasterized energy field. Rasterizing the damage energy field facilitates efficient damage performance assessment in subsequent steps.

[0011] The specific implementation method of step one is as follows:

[0012] Determine the maximum damage range of the aircraft, i.e., width W. p meters and length L p Meters divide the energy field into G square grids K, each with a side length of b meters. rt (r=1,…,r p ;t=1,…,t p The coordinates of the center point of each square are: Vertex coordinates are Define W p and L p It is an integer multiple of b, that is, r p =W p / b,t p =L p / b,r p ×t p =G, which makes the gridded energy field more regular.

[0013] Calculate the energy field information contained in each square. For the fragment energy field, at the descent velocity v of the spacecraft... j Angle of landing θ j Height z j The fragments produced by the aircraft under these circumstances are as follows:

[0014]

[0015] in, and These represent the horizontal and vertical coordinates, mass, and velocity of the fragment, respectively.

[0016] Based on the rasterized grid distribution, the number of fragments in each grid was calculated as follows:

[0017]

[0018] Among them, E f This represents the kinetic energy threshold required to penetrate the target. Based on equation (2), the fragment energy field in equation (1) is transformed into a gridded energy field, as shown below:

[0019]

[0020] Step two involves constructing a diffusion model for rapid energy field prediction. This model comprises an input layer, two linear layers, five U-Net modules, and an output layer. The input layer calculates the spacecraft's descent velocity v. j Angle of landing θ j Height z j Three information inputs are fed into the linear layers of the model. Two linear layers map the three inputs to feature maps in a high-dimensional feature space. By using linear layers, the traditional diffusion model avoids its dependence on image input, enabling it to more flexibly address energy field prediction problems. The U-Net block treats the feature maps obtained from the linear layers as "pure noise," and performs denoising processing step by step through five U-Net blocks to generate a matrix representing the characteristics of the energy field.

[0021] During the training data generation phase, input samples are obtained using Latin hypercube sampling, based on the range of values ​​for the aircraft's landing velocity, landing angle, and altitude. The energy field is then calculated using a real numerical simulation model based on the input samples. This energy field is then rasterized in step one to obtain the output samples. Mean squared error is used as the loss function during the training of the diffusion model.

[0022] Step three involves constructing a deterministic damage assessment model based on the rasterized energy field. This model determines that all fragments within a grid fall on the target by ensuring the grid center is on the target, thus accelerating the damage assessment process while maintaining accuracy. Building upon the deterministic damage assessment model, a rapid damage expectation assessment model is established, considering both CEP and TEP, and incorporating chaotic polynomials. The chaotic polynomial coefficients are calculated using the projection method, and the damage expectation is then calculated based on these coefficients.

[0023] The specific implementation method for step three is as follows:

[0024] Under certain conditions, the damage effect caused by the aircraft to the target is represented by the damage degree d, given the aiming point (x). j ,y j ) and target coordinates (x) t ,y t Assume the target consists of S key components, each with a weight of s. i The formula for calculating the degree of damage d caused by the aircraft to the target is as follows:

[0025]

[0026] Among them, G i (x j ,y j E f The damage law for the i-th component of the target is represented by ), and it is calculated as follows:

[0027]

[0028] Among them, E f S represents the kinetic energy threshold required to penetrate a target. i (x j ,y j E f N represents the number of fragments that fall on the i-th component and satisfy the kinetic energy threshold. si This represents the threshold number of fragments required to destroy the component. In calculating S... i (x j ,y j E f When generating the predicted energy field based on step two, this energy field is rasterized, eliminating the need to calculate whether each fragment falls on the target component individually; only the coordinates of the grid center need to be calculated. If a fragment falls on a component, it is determined that all fragments in that square have fallen on the component, thus accelerating the efficiency of damage assessment.

[0029] After considering the uncertainties of CEP and TEP, the damage caused by the aircraft to the target will fluctuate. It is necessary to calculate the expected damage D of the aircraft to the target, also known as the damage probability D, given a set of aiming points (x1, y1, x2, y2, ..., x...). a ,y a and target coordinates (x) t ,y t Let d be a second-order chaotic polynomial model, with random input η = [x] t ,y t ,x1,y1,…,x a ,y a ]:

[0030]

[0031] Where ξ is a standard random variable, and it has a linear transformation relationship with the random input η, Φ i (ξ) is a multivariate orthogonal polynomial, b i represents the coefficients of the chaotic polynomial, and represents the quantity to be estimated.

[0032] b i The calculation is performed using the projection method, as follows:

[0033]

[0034] In the formula,

[0035] It expresses expectation.

[0036] Complete b i After calculation, the chaotic polynomial model is constructed. The formula for calculating the expected damage D considering the effects of CEP and TEP is as follows:

[0037]

[0038] Step four: Considering that the energy field usually has axisymmetric characteristics, in the XOY coordinate system, equidistant aiming points are generated on the X and Y axes at a distance of B. At each aiming point, the method in step three is applied to calculate the expected damage value, and the maximum expected damage value is selected as the approximate maximum expected damage value of the aircraft to the target. This maximum expected damage value is an important parameter for firepower allocation.

[0039] Step 5: Select the maximum damage area in the current aircraft as the target grouping criterion, denoted as J. Sort the targets according to their damage determination from high to low, and assign the first target to the first group. Calculate the distance between the remaining targets and the first target. If the distance is less than J, the target is assigned to the first group. Denote the first grouping as T1. Then, sort the remaining targets according to their damage determination from high to low, and assign the target with the highest damage determination to the second group. Repeat the same steps to complete the second grouping, denoted as T2. Repeat the above grouping process until all targets have been grouped, resulting in k groups T1, T2, ..., T k .

[0040] Step Six: For fire allocation under unsaturated conditions, the allocation combination of aircraft and targets is used as design variables. The maximum damage effectiveness, minimum damage deviation, and minimum number of aircraft used are selected as optimization objectives. The constraint functions include: ensuring that the expected damage caused by the final optimization result is large enough and that the optimization result fits the damage determination, while reducing mission costs; and abandoning some unimportant targets by dynamically adjusting the target damage determination, thereby ensuring that important targets are damaged to a limited extent. Based on the above design variables, objective functions, and constraint functions, the fire allocation optimization model under unsaturated mission conditions is obtained.

[0041] The specific implementation method for step six is ​​as follows:

[0042] Step 6.1: Using the allocation and combination of aircraft and targets as design variables, with a total of m targets and n aircraft, the aircraft and targets form the following combinations:

[0043]

[0044] Where: z ji ∈{0,1} represents the combination of aircraft j and target i (i=1,2,…m,j=1,2,n), and is the design variable for firepower allocation; z ji =0 means that aircraft j is not assigned to target i, z ji =1 indicates allocation.

[0045] Step 6.2: Select the maximum damage effectiveness as one of the optimization objectives to ensure that the expected damage caused by the final optimization result is large enough; select the minimum damage deviation as one of the optimization objectives to ensure that the optimization result fits the damage determination; select the minimum number of aircraft used as one of the optimization objectives to reduce mission costs.

[0046] The maximum damage effectiveness is calculated as follows:

[0047]

[0048] Among them, D i D represents the expected total damage to this group of targets. ji This represents the expected damage, whether direct or incidental, caused by aircraft j to target i. This indicates the determination to damage target i.

[0049] If z ji =1, R ji The definition is as follows:

[0050]

[0051] Among them, L ji This indicates the distance between the launch platform of aircraft j and the target. and Let J represent the minimum and maximum ranges of aircraft J, respectively.

[0052] If z ij =0, and aircraft j is assigned to group T a Internal target q (q=1,2,…,i-1,i+1,…,m) a ), that is, z jq If the value is 1, then aircraft j will cause collateral damage to target i. The collateral damage effects are as follows:

[0053]

[0054] in, The distance between targets i and q is defined by L1, L2, α, γ, and β, which are manually specified.

[0055] Minimizing the deviation, while satisfying the damage determination, avoids excessive damage as much as possible, and is calculated using the following formula:

[0056]

[0057] Among them, D i This represents the actual damage effect on target i under the current firepower allocation.

[0058] According to equation (15), minimize the number of aircraft used to avoid resource waste:

[0059]

[0060] Step 6.3: For unsaturated mission scenarios, to ensure that important targets with high damage determination meet the damage requirements, a constraint function for dynamically adjusting the target damage determination is adopted to constrain the remaining targets to meet the damage determination, while abandoning some unimportant targets. The maximum number of aircraft required for each target is obtained through equation (16):

[0061]

[0062] For non-saturated tasks, it is obvious that Starting with the target that minimizes damage, set the maximum number of aircraft for each of them to zero (i.e., ...). If the updated number of aircraft remains the same... Then set the maximum number of aircraft for the target with the second smallest damage determination to 0, and repeat this process until the condition is met. Up to this point. Based on the above calculations, update the damage determination constraint:

[0063]

[0064] Based on the above design variables, optimization objectives, and constraint functions, the firepower allocation model is constructed as follows:

[0065]

[0066] Step 7: Reduce the high-dimensional design variables of the aircraft-target allocation combination to a one-dimensional integer vector to improve the efficiency of the fire allocation model optimization. In the process of optimizing the fire allocation model, the fire allocation optimization model is equivalent to a saturated task. A single-target optimization model for fire allocation under saturation is constructed and solved. If the solution satisfies the constraints of damage determination for all targets, the optimization ends and the aircraft-target allocation result is obtained. If the solution cannot satisfy the constraints of damage determination for all targets, the fire allocation optimization model is equivalent to an unsaturated task. A fire allocation model under unsaturated task is constructed and optimized, and the optimized aircraft-target allocation result is obtained.

[0067] The specific implementation method for step seven is as follows:

[0068] To ensure optimization efficiency, the design variables are transformed into a one-dimensional integer vector Q = [q1, ..., q...]. j ,…,q n ], where 0≤q j ≤m and q j =i means assigning aircraft j to target i, if q j =0 indicates that the aircraft was not used.

[0069] For saturation missions, the destruction of all targets is inevitable; therefore, the focus should be on controlling the damage deviation to zero and minimizing the number of aircraft used. The fire allocation optimization model is constructed as follows:

[0070]

[0071] After the firepower allocation optimization model is optimized, it is necessary to determine whether the current allocation scheme meets the damage requirements of each target, i.e. (i = 1, 2, ..., m). If the damage requirement for each target is not met, it indicates an unsaturated mission, requiring the addition of dynamic adjustment of damage determination constraints. Unsaturated missions do not require control over the number of aircraft used; while ensuring damage to important targets, they may damage multiple targets with high damage requirements. Therefore, while ensuring the damage index f1 is as large as possible, the deviation f2 should be reduced. The firepower allocation model is as follows:

[0072]

[0073] Solving equation (20) yields the aircraft-target allocation result.

[0074] Step 8: Based on the target grouping situation in Step 6 and the aircraft-target allocation results in Step 7, the fire planning problem is broken down into several intra-group aiming point planning problems with groups as the unit, and then the final fire planning result of hierarchical fire planning is obtained by solving them.

[0075] The specific implementation method for step eight is as follows:

[0076] Based on the dynamically adjusted damage determination constraints constructed in step seven Determine the damage requirements for this group of targets. The objective of aiming point allocation is to maximize the damage effect while satisfying the damage requirements. Aiming point optimization is performed in groups, therefore the optimization model is as follows:

[0077]

[0078] Among them, D i The damage effect under the current aiming point scheme; x L x U y L and y U This refers to the range of the aiming point.

[0079] The final hierarchical firepower planning result is obtained by solving equation (21).

[0080] It also includes step nine: based on the final hierarchical fire planning results obtained in step eight, accurately identify the importance and threat level of different targets, accurately assess the actual damage caused by the aircraft to the targets by considering CEP and TEP, match the mission planning results with the actual results, and ensure that the aircraft aiming point can be obtained quickly in the real mission environment by optimizing the fire planning model, so as to achieve high real-time performance of the aircraft.

[0081] Beneficial effects:

[0082] 1. The present invention discloses a hierarchical firepower planning method based on a fusion diffusion model and a chaotic polynomial. It is based on a gridded preprocessing method with the maximum range of energy fields, and adopts a square gridded method and statistically analyzes the energy field information contained in each grid, which helps to achieve efficient damage performance assessment.

[0083] 2. This invention discloses a hierarchical fire control planning method that integrates a diffusion model and a chaotic polynomial, constructing a diffusion model for rapid energy field prediction. The diffusion model includes an input layer, two linear layers, five U-Net modules, and an output layer. The input layer calculates the aircraft's descent velocity v. j Angle of landing θ j Height z jThree information inputs are fed into the model's linear layers, and two linear layers map these three inputs to feature maps in a high-dimensional feature space. The linear layers avoid the dependence of traditional diffusion models on image input, allowing the diffusion model to more flexibly address energy field prediction problems. The U-Net blocks treat the feature maps obtained from the linear layers as "pure noise," and perform progressive denoising through five U-Net blocks to generate a matrix representing the energy field characteristics. Energy field prediction based on the diffusion model, by mapping the aircraft state parameters to high-dimensional feature maps and then progressively denoising to restore energy field details, accelerates energy field prediction efficiency while maintaining prediction accuracy, facilitating rapid loss assessment.

[0084] 3. This invention discloses a fusion diffusion model and a hierarchical fire planning method based on chaotic polynomials. Based on a gridded energy field, a damage assessment model under deterministic conditions is constructed. By determining whether the center of a square in the energy field falls on the target, it is determined that all fragments within the square fall on the target, thus ensuring the accuracy of damage assessment while accelerating its efficiency. Building upon the deterministic damage assessment model, a rapid damage expectation assessment method based on chaotic polynomials is constructed. This method fully considers the uncertainties of the circular probability error (CEP) of the aircraft and the target position detection error (TEP), and uses a projection method to calculate the chaotic polynomial coefficients, significantly improving the prediction efficiency of damage expectation.

[0085] 4. The hierarchical fire planning method of fusion diffusion model and chaotic polynomial disclosed in this invention adopts a target grouping strategy that considers collateral damage. By using the maximum range of the aircraft's energy field as a criterion, targets within the criterion are grouped together, which effectively improves resource utilization.

[0086] 5. The present invention discloses a hierarchical fire planning method based on a fusion diffusion model and a chaotic polynomial, which considers the fire allocation results of unsaturated tasks, dynamically adjusts damage determination constraints in a general manner, ensures the priority of important targets, and improves the adaptability of fire allocation in complex environments.

[0087] 6. The present invention discloses a hierarchical fire planning method based on a fusion diffusion model and a chaotic polynomial. The hierarchical fire planning method is constructed, and a hierarchical fire planning strategy is established with a top-level target group-based fire allocation and a bottom-level aiming point planning for each group. This significantly improves the aiming point generation speed and ensures that the aircraft aiming point can be obtained quickly in a real-time mission environment, thereby achieving high real-time performance of the aircraft. Attached Figure Description

[0088] Figure 1 A flowchart of a hierarchical firepower planning method that integrates a diffusion model and chaotic polynomials;

[0089] Figure 2 This is a schematic diagram of a gridded energy field;

[0090] Figure 3 To improve the network structure of the diffusion model;

[0091] Figure 4 The deviation between the energy field predicted by the diffusion model and the actual energy field;

[0092] Figure 5 This is a diagram illustrating the grouping strategy.

[0093] Figure 6 Results of tiered firepower planning. Detailed Implementation

[0094] To better illustrate the purpose and advantages of the present invention, the invention will be further described below in conjunction with the accompanying drawings and examples.

[0095] Example 1:

[0096] like Figure 1 As shown in the figure, the hierarchical firepower planning method of fusion diffusion model and chaotic polynomial disclosed in this embodiment is implemented in the following specific steps:

[0097] Step one involves dividing the energy field into several square grids according to the length and width of the damage range of the aircraft's energy field, with predetermined side lengths. The energy field information contained within each grid is then statistically analyzed. For fragment energy fields, the number of fragments meeting the kinetic energy threshold within each grid is counted, resulting in a rasterized energy field. Rasterizing the damage energy field facilitates efficient damage performance assessment in subsequent steps.

[0098] The specific implementation method of step one is as follows:

[0099] like Figure 2 As shown, the maximum damage range of the aircraft is determined, i.e., the width W. p meters and length L p Meters divide the energy field into G square grids K, each with a side length of b meters. rt (r=1,…,r p ;t=1,…,t p The coordinates of the center point of each square are: Vertex coordinates are Define W p and L p It is an integer multiple of b, that is, r p =W p / b,t p =L p / b,r p ×t p =G, which makes the gridded energy field more regular.

[0100] Calculate the energy field information contained in each square. For the fragment energy field, at the descent velocity v of the spacecraft... j Angle of landing θ j Height z j The fragments produced by the aircraft under these circumstances are as follows:

[0101]

[0102] in, and These represent the horizontal and vertical coordinates, mass, and velocity of the fragment, respectively.

[0103] Based on the rasterized grid distribution, the number of fragments in each grid was calculated as follows:

[0104]

[0105] Among them, E f This represents the kinetic energy threshold required to penetrate the target. Based on equation (2), the fragment energy field in equation (1) is transformed into a gridded energy field, as shown below:

[0106]

[0107] Step two: Construct a diffusion model for rapid energy field prediction. The network structure of the constructed diffusion model is as follows: Figure 3 As shown, the diffusion model consists of an input layer, two linear layers, five U-Net modules, and an output layer. The input layer calculates the aircraft's descent velocity v. j Angle of landing θ j Height z j Three information inputs are fed into the linear layers of the model. Two linear layers map the three inputs to feature maps in a high-dimensional feature space. By using linear layers, the traditional diffusion model avoids its dependence on image input, enabling it to more flexibly address energy field prediction problems. The U-Net block treats the feature maps obtained from the linear layers as "pure noise," and performs denoising processing step by step through five U-Net blocks to generate a matrix representing the characteristics of the energy field.

[0108] During the training data generation phase, input samples are obtained using Latin hypercube sampling, based on the range of values ​​for the aircraft's landing velocity, landing angle, and altitude. The energy field is then calculated using a real numerical simulation model based on the input samples. This energy field is then rasterized in step one to obtain the output samples. Mean squared error is used as the loss function during the training of the diffusion model.

[0109] Figure 4 The prediction bias of the energy field is shown under the conditions of velocity v = 700 m / s, angle θ = 90°, and height z = 25 m. Figure 4 It can be seen that the diffusion model has high accuracy in predicting the kinetic energy of each grid, and the deviation between the predicted results and the actual values ​​is small, with a maximum deviation of only about 12, which fully verifies the reliability and accuracy of the model in complex flow field simulation.

[0110] Step three involves constructing a deterministic damage assessment model based on the rasterized energy field. This model determines that all fragments within a grid fall on the target by ensuring the grid center is on the target, thus accelerating the damage assessment process while maintaining accuracy. Building upon the deterministic damage assessment model, a rapid damage expectation assessment model is established, considering both CEP and TEP, and incorporating chaotic polynomials. The chaotic polynomial coefficients are calculated using the projection method, and the damage expectation is then calculated based on these coefficients.

[0111] The specific implementation method for step three is as follows:

[0112] Under certain conditions, the damage effect caused by the aircraft to the target is represented by the damage degree d, given the aiming point (x). j ,y j ) and target coordinates (x) t ,y t Assume the target consists of S key components, each with a weight of s. i The formula for calculating the degree of damage d caused by the aircraft to the target is as follows:

[0113]

[0114] Among them, G i (x j ,y j E f The damage law for the i-th component of the target is represented by ), and it is calculated as follows:

[0115]

[0116] Among them, E f S represents the kinetic energy threshold required to penetrate a target. i (x j ,y j E f N represents the number of fragments that fall on the i-th component and satisfy the kinetic energy threshold. si This represents the threshold number of fragments required to destroy the component. In calculating S... i (x j ,y j E f When generating the predicted energy field based on step two, this energy field is rasterized, eliminating the need to calculate whether each fragment falls on the target component individually; only the coordinates of the grid center need to be calculated. If a fragment falls on a component, it is determined that all fragments in that square have fallen on the component, thus accelerating the efficiency of damage assessment.

[0117] After considering the uncertainties of CEP and TEP, the damage caused by the aircraft to the target will fluctuate. It is necessary to calculate the expected damage D of the aircraft to the target, also known as the damage probability D, given a set of aiming points (x1, y1, x2, y2, ..., x...). a ,y a and target coordinates (x) t ,y t Let d be a second-order chaotic polynomial model, with random input η = [x] t ,y t ,x1,y1,…,x a ,y a ]:

[0118]

[0119] Where ξ is a standard random variable, and it has a linear transformation relationship with the random input η, Φ i (ξ) is a multivariate orthogonal polynomial, b i represents the coefficients of the chaotic polynomial, and represents the quantity to be estimated.

[0120] b i The calculation is performed using the projection method, as follows:

[0121]

[0122] In the formula,

[0123] It expresses expectation.

[0124] Complete b i After calculation, the chaotic polynomial model is constructed. The formula for calculating the expected damage D considering the effects of CEP and TEP is as follows:

[0125]

[0126] Step four: Considering that the energy field usually has axisymmetric characteristics, in the XOY coordinate system, equidistant aiming points are generated on the X and Y axes at a distance of B. At each aiming point, the method in step three is applied to calculate the expected damage value, and the maximum expected damage value is selected as the approximate maximum expected damage value of the aircraft to the target. This maximum expected damage value is an important parameter for firepower allocation.

[0127] Step 5, the proposed grouping strategy is as follows: Figure 5As shown, the maximum damage area in the current aircraft is selected as the target grouping criterion, denoted as J. Targets are sorted from highest to lowest damage determination, and the target ranked first is assigned to group 1. The distance between each of the remaining targets and this target is calculated; if the distance is less than J, the target is assigned to group 1. This first grouping is denoted as T1. Then, the remaining targets are sorted from highest to lowest damage determination, and the target with the highest damage determination is assigned to group 2. The same steps are repeated to complete the second grouping, denoted as T2. This grouping process is repeated until all targets have been grouped, resulting in k groups T1, T2, ..., T k .

[0128] Step Six: For fire allocation under unsaturated conditions, the allocation combination of aircraft and targets is used as design variables. The maximum damage effectiveness, minimum damage deviation, and minimum number of aircraft used are selected as optimization objectives. The constraint functions include: ensuring that the expected damage caused by the final optimization result is large enough and that the optimization result fits the damage determination, while reducing mission costs; and abandoning some unimportant targets by dynamically adjusting the target damage determination, thereby ensuring that important targets are damaged to a limited extent. Based on the above design variables, objective functions, and constraint functions, the fire allocation optimization model under unsaturated mission conditions is obtained.

[0129] The specific implementation method for step six is ​​as follows:

[0130] Step 6.1: Using the allocation and combination of aircraft and targets as design variables, with a total of m targets and n aircraft, the aircraft and targets form the following combinations:

[0131]

[0132] Where: z ji ∈{0,1} represents the combination of aircraft j and target i (i=1,2,…m,j=1,2,n), and is the design variable for firepower allocation; z ji =0 means that aircraft j is not assigned to target i, z ji =1 indicates allocation.

[0133] Step 6.2: Select the maximum damage effectiveness as one of the optimization objectives to ensure that the expected damage caused by the final optimization result is large enough; select the minimum damage deviation as one of the optimization objectives to ensure that the optimization result fits the damage determination; select the minimum number of aircraft used as one of the optimization objectives to reduce mission costs.

[0134] The maximum damage effectiveness is calculated as follows:

[0135]

[0136] Among them, D iD represents the expected total damage to this group of targets. ji This represents the expected damage, whether direct or incidental, caused by aircraft j to target i. This indicates the determination to damage target i.

[0137] If z ji =1, R ji The definition is as follows:

[0138]

[0139] Among them, L ji This indicates the distance between the launch platform of aircraft j and the target. and Let J represent the minimum and maximum ranges of aircraft J, respectively.

[0140] If z ij =0, and aircraft j is assigned to group T a Internal target q (q=1,2,…,i-1,i+1,…,m) a ), that is, z jq If the value is 1, then aircraft j will cause collateral damage to target i. The collateral damage effects are as follows:

[0141]

[0142] in, The distance between targets i and q is defined by L1, L2, α, γ, and β, which are manually specified.

[0143] Minimizing the deviation, while satisfying the damage determination, avoids excessive damage as much as possible, and is calculated using the following formula:

[0144]

[0145] Among them, D i This represents the actual damage effect on target i under the current firepower allocation.

[0146] According to equation (15), minimize the number of aircraft used to avoid resource waste:

[0147]

[0148] Step 6.3: For unsaturated mission scenarios, to ensure that important targets with high damage determination meet the damage requirements, a constraint function for dynamically adjusting the target damage determination is adopted to constrain the remaining targets to meet the damage determination, while abandoning some unimportant targets. The maximum number of aircraft required for each target is obtained through equation (16):

[0149]

[0150] For non-saturated tasks, it is obvious that Starting with the target that minimizes damage, set the maximum number of aircraft for each of them to zero (i.e., ...). If the updated number of aircraft remains the same... Then set the maximum number of aircraft for the target with the second smallest damage determination to 0, and repeat this process until the condition is met. Up to this point. Based on the above calculations, update the damage determination constraint:

[0151]

[0152] Based on the above design variables, optimization objectives, and constraint functions, the firepower allocation model is constructed as follows:

[0153]

[0154] Step 7: Reduce the high-dimensional design variables of the aircraft-target allocation combination to a one-dimensional integer vector to improve the efficiency of the fire allocation model optimization. In the process of optimizing the fire allocation model, the fire allocation optimization model is equivalent to a saturated task. A single-target optimization model for fire allocation under saturation is constructed and solved. If the solution satisfies the constraints of damage determination for all targets, the optimization ends and the aircraft-target allocation result is obtained. If the solution cannot satisfy the constraints of damage determination for all targets, the fire allocation optimization model is equivalent to an unsaturated task. A fire allocation model under unsaturated task is constructed and optimized, and the optimized aircraft-target allocation result is obtained.

[0155] The specific implementation method for step seven is as follows:

[0156] To ensure optimization efficiency, the design variables are transformed into a one-dimensional integer vector Q = [q1, ..., q...]. j ,…,q n ], where 0≤q j ≤m and q j =i means assigning aircraft j to target i, if q j =0 indicates that the aircraft was not used.

[0157] For saturation missions, the destruction of all targets is inevitable; therefore, the focus should be on controlling the damage deviation to zero and minimizing the number of aircraft used. The fire allocation optimization model is constructed as follows:

[0158]

[0159] After the firepower allocation optimization model is optimized, it is necessary to determine whether the current allocation scheme meets the damage requirements of each target, i.e. If the damage requirements for each target are not met, it indicates an unsaturated mission, requiring the addition of dynamic damage determination constraints. Unsaturated missions do not require control over the number of aircraft used; while ensuring damage to critical targets, they may damage multiple targets with high damage requirements. Therefore, while maximizing the damage index f1, the deviation f2 should be reduced. The firepower allocation model is as follows:

[0160]

[0161] Solving equation (20) yields the aircraft-target allocation result.

[0162] Step 8: Based on the target grouping situation in Step 6 and the aircraft-target allocation results in Step 7, the fire planning problem is broken down into several intra-group aiming point planning problems with groups as the unit, and then the final fire planning result of hierarchical fire planning is obtained by solving them.

[0163] The specific implementation method for step eight is as follows:

[0164] Based on the dynamically adjusted damage determination constraints constructed in step seven Determine the damage requirements for this group of targets. The objective of aiming point allocation is to maximize the damage effect while satisfying the damage requirements. Aiming point optimization is performed in groups, therefore the optimization model is as follows:

[0165]

[0166] Among them, D i The damage effect under the current aiming point scheme; x L x U y L and y U This refers to the range of the aiming point.

[0167] The final hierarchical firepower planning result is obtained by solving equation (21).

[0168] Figure 6 The results of the vitality planning are shown when CEP = TEP = 10m and the number of targets is 3. The red pentagrams in the figure represent the coordinates of the aiming point, which are obtained by solving equation (21). The red ring is a simplified schematic diagram of the aircraft's energy field, which is its main damage range. The black squares in the figure represent the mission targets. As can be seen from the figure, the optimized results fully consider the impact of collateral damage. The mission requirements were completed using only 2 aircraft, which greatly improved the utilization efficiency of the aircraft.

[0169] Step Nine: Based on the final hierarchical fire planning results obtained in Step Eight, accurately identify the importance and threat level of different targets. By considering CEP and TEP, accurately assess the actual damage caused by the aircraft to the targets, so that the mission planning results match the actual results. By optimizing the fire planning model, ensure that the aircraft aiming point can be obtained quickly in the real-time mission environment, and achieve high real-time performance of the aircraft.

[0170] The above detailed description further illustrates the purpose, technical solution, and beneficial effects of the invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A hierarchical firepower planning method that integrates diffusion models and chaotic polynomials, characterized in that: A target damage expectation assessment method is established, which includes unified matrix representation of energy field rasterization, energy field prediction using diffusion model, and damage expectation calculation using chaotic polynomial. Based on the rasterization preprocessing method with the maximum range of energy field, the energy field information contained in each grid is statistically analyzed. A hierarchical optimization fire planning architecture is established. The top layer determines the allocation matrix between aircraft and targets for fire allocation. At the same time, a grouping strategy is used to consider the collateral damage effect of aircraft on targets. Based on this, the bottom layer optimizes the aiming point for each group using a fast and accurate target damage expectation assessment method, quickly obtains the aiming point planning results, and realizes the high real-time performance of aircraft based on the aiming point planning results.

2. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 1, characterized in that: Includes the following steps, Step 1: Based on the length and width of the damage range of the spacecraft's energy field, divide the energy field into several square grids according to the predetermined side length, count the energy field information contained in each grid, and for the fragment energy field, count the number of fragments in each grid that meet the kinetic energy threshold to obtain the gridded energy field. Step two: Construct a diffusion model for rapid energy field prediction. The diffusion model includes an input layer, two linear layers, five U-Net modules, and an output layer; the input layer will input the spacecraft's descent velocity v. j Angle of landing θ j Height z j Three information inputs are fed into the linear layers of the model. Two linear layers map the three input information to feature maps in a high-dimensional feature space. The linear layers avoid dependence on image input. The U-Net block treats the feature maps obtained by the linear layers as "pure noise" and performs denoising processing step by step through five U-Net blocks to generate a matrix representing the energy field characteristics. Step 3: Based on the rasterized energy field, construct a damage assessment model under deterministic conditions. By determining whether the center of the square in the energy field falls on the target, it is determined that all fragments in the square fall on the target, thereby ensuring the accuracy of damage assessment while accelerating the efficiency of damage assessment. On the basis of the deterministic damage assessment model, considering CEP and TEP, and combining chaotic polynomials, establish a rapid damage expectation assessment model. The chaotic polynomial coefficients are calculated by the projection method, and the damage expectation is calculated based on the polynomial coefficients. Step four: Considering that the energy field usually has axisymmetric properties, in the XOY coordinate system, equidistant aiming points are generated on the X and Y axes at a distance of B. At each aiming point, the method in step three is applied to calculate the expected damage value, and the maximum expected damage value is selected as the approximate maximum expected damage value of the aircraft to the target. Step 5: Select the maximum damage area in the current aircraft as the target grouping criterion, denoted as J. Sort the targets according to their damage determination from high to low, and assign the first target to the first group. Calculate the distance between the remaining targets and the first target. If the distance is less than J, the target is assigned to the first group. Denote the first grouping as T1. Then, sort the remaining targets according to their damage determination from high to low, and assign the target with the highest damage determination to the second group. Repeat the same steps to complete the second grouping, denoted as T2. Repeat the above grouping process until all targets have been grouped, resulting in k groups T1, T2, ..., T k ; Step 6: For firepower allocation under unsaturated conditions, the allocation combination of aircraft and targets is used as design variables, and the maximum damage effectiveness, minimum damage deviation, and minimum number of aircraft used are selected as optimization objectives. The constraint functions include: ensuring that the expected damage caused by the final optimization result is large enough, and that the optimization result fits the damage determination, while reducing the task cost; By dynamically adjusting the target damage determination to abandon some unimportant targets, the important targets are guaranteed to be destroyed to a limited extent. Based on the above design variables, objective function, and constraint function, a firepower allocation optimization model for unsaturated mission conditions is obtained. Step 7: Reduce the high-dimensional design variables of the aircraft-target allocation combination to a one-dimensional integer vector; in the process of optimizing the fire allocation optimization model, the fire allocation optimization model is equivalent to a saturated task, and a single-target optimization model for fire allocation under saturation is constructed and solved. If the solution result satisfies the constraints of damage determination of all targets, the optimization ends and the aircraft-target allocation result is obtained; if the solution result cannot satisfy the constraints of damage determination of all targets, the fire allocation optimization model is equivalent to an unsaturated task, and a fire allocation model under unsaturated task is constructed and optimized, resulting in the aircraft-target allocation result after optimization. Step 8: Based on the target grouping situation in Step 6 and the aircraft-target allocation results in Step 7, the fire planning problem is broken down into several intra-group aiming point planning problems with groups as the unit, and then the final fire planning result of hierarchical fire planning is obtained by solving them.

3. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 2, characterized in that: The specific implementation method of step one is as follows: Determine the maximum damage range of the aircraft, i.e., width W. p meters and length L p Meters divide the energy field into G square grids K, each with a side length of b meters. rt (r=1,…,r p ;t=1,…,t p The coordinates of the center point of each square are... Vertex coordinates are Define W p and L p It is an integer multiple of b, that is, r p =W p / b,t p =L p / b,r p ×t p =G, making the gridded energy field more regular; Calculate the energy field information contained in each square. For the fragment energy field, at the descent velocity v of the spacecraft... j Angle of landing θ j Height z j The fragments produced by the aircraft under these circumstances are as follows: in, and These represent the horizontal and vertical coordinates, mass, and velocity of the fragment, respectively. Based on the rasterized grid distribution, the number of fragments in each grid was calculated as follows: Among them, E f This represents the kinetic energy threshold required to penetrate the target; according to equation (2), the fragment energy field in equation (1) is converted into a gridded energy field, and the gridded energy field is as follows:

4. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 3, characterized in that: In step two, In the training data generation phase, input samples are obtained using Latin hypercube sampling based on the range of values ​​for the aircraft's landing speed, landing angle, and altitude. The energy field is calculated using a real numerical simulation model based on the input samples, and the energy field is rasterized in step one to obtain the output samples. The mean square error is used as the loss function during the training of the diffusion model.

5. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 4, characterized in that: The specific implementation method of step three is as follows: Under certain conditions, the damage effect caused by the aircraft to the target is represented by the damage degree d, given the aiming point (x). j ,y j ) and target coordinates (x) t ,y t Assume the target consists of S key components, each with a weight of s. i The formula for calculating the degree of damage d caused by the aircraft to the target is as follows: Among them, G i (x j ,y j E f The damage law for the i-th component of the target is represented by ), and it is calculated as follows: Among them, E f S represents the kinetic energy threshold required to penetrate a target. i (x j ,y j E f N represents the number of fragments that fall on the i-th component and satisfy the kinetic energy threshold. si This represents the threshold number of fragments required to destroy the component; in calculating S... i (x j ,y j E f When generating the predicted energy field based on step two, this energy field is rasterized, eliminating the need to calculate whether each fragment falls on the target component individually; only the coordinates of the grid center need to be calculated. Whether it falls on a fragment; if it falls on a fragment, then all fragments in that square are determined to have fallen on the component, thus speeding up the efficiency of damage effect assessment. After considering the uncertainties of CEP and TEP, the damage caused by the aircraft to the target will fluctuate. It is necessary to calculate the expected damage D of the aircraft to the target, also known as the damage probability D, given a set of aiming points (x1, y1, x2, y2, ..., x...). a ,y a and target coordinates (x) t ,y t Let d be a second-order chaotic polynomial model, with random input η = [x] t ,y t ,x1,y1,…,x a ,y a ]: Where ξ is a standard random variable, and it has a linear transformation relationship with the random input η, Φ i (ξ) is a multivariate orthogonal polynomial, b i Let be the coefficients of the chaotic polynomial, and be the quantity to be estimated; b i The calculation is performed using the projection method, as follows: In the formula, Expressing expectations; Complete b i After calculation, the chaotic polynomial model is constructed. The formula for calculating the expected damage D considering the effects of CEP and TEP is as follows:

6. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 5, characterized in that: The specific implementation method for step six is ​​as follows: Step 6.1: Using the allocation and combination of aircraft and targets as design variables, with a total of m targets and n aircraft, the aircraft and targets form the following combinations: Where: z ji ∈{0,1} represents the combination of aircraft j and target i (i=1,2,…m,j=1,2,n), and is the design variable for firepower allocation; z ji =0 means that aircraft j is not assigned to target i, z ji =1 indicates allocation; Step 6.2: Select the maximum damage effectiveness as one of the optimization objectives to ensure that the expected damage caused by the final optimization result is large enough; select the minimum damage deviation as one of the optimization objectives to ensure that the optimization result fits the damage determination; select the minimum number of aircraft used as one of the optimization objectives to reduce mission costs; The maximum damage effectiveness is calculated as follows: Among them, D i D represents the expected total damage to this group of targets. ji This represents the expected damage, whether direct or incidental, caused by aircraft j to target i. This indicates the determination to harm target i; If z ji =1, R ji The definition is as follows: Among them, L ji This indicates the distance between the launch platform of aircraft j and the target. and Let represent the minimum and maximum ranges of aircraft j, respectively; If z ij =0, and aircraft j is assigned to group T a Internal target q (q=1,2,…,i-1,i+1,…,m) a ), that is, z jq If the value is 1, then aircraft j will cause collateral damage to target i. The collateral damage effects are as follows: in, The distance between targets i and q, L1, L2, α, γ, and β are specified manually; Minimizing the deviation, while satisfying the damage determination, avoids excessive damage as much as possible, and is calculated using the following formula: Among them, D i This represents the actual damage effect on target i under the current firepower allocation result; According to equation (15), minimize the number of aircraft used to avoid resource waste: Step 6.3: For unsaturated mission scenarios, to ensure that important targets with high damage determination meet the damage requirements, a constraint function for dynamically adjusting the target damage determination is adopted to constrain the remaining targets to meet the damage determination, while abandoning some unimportant targets; the maximum number of aircraft required for each target is obtained through equation (16): For non-saturated tasks, it is obvious that Starting with the target that minimizes damage, set the maximum number of aircraft for each of them to zero (i.e., ...). If the updated number of aircraft remains the same... Then set the maximum number of aircraft for the target with the second smallest damage determination to 0, and repeat this process until the condition is met. Up to this point; update the damage determination constraint based on the above calculations: Based on the above design variables, optimization objectives, and constraint functions, the firepower allocation model is constructed as follows:

7. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 6, characterized in that: The specific implementation method for step seven is as follows: To ensure optimization efficiency, the design variables are transformed into a one-dimensional integer vector Q = [q1, ..., q...]. j ,…,q n ], where 0≤q j ≤m and q j ∈Z, q j =i means assigning aircraft j to target i, if q j =0 indicates that the aircraft was not used; For saturation missions, the destruction of all targets is inevitable; therefore, it is sufficient to control the damage deviation to zero and minimize the number of aircraft used. The fire allocation optimization model is constructed as follows: After the firepower allocation optimization model is optimized, it is necessary to determine whether the current allocation scheme meets the damage requirements of each target, i.e. If the damage requirements for each target are not met, it indicates an unsaturated mission, requiring the addition of dynamic damage determination constraints. Unsaturated missions do not require control over the number of aircraft used; while ensuring damage to critical targets, they may damage multiple targets with high damage requirements. Therefore, while maximizing the damage index f1, the deviation f2 should be reduced. The firepower allocation model is as follows: Solve equation (20) to obtain the aircraft-target allocation result.

8. The hierarchical firepower planning method of fusion diffusion model and chaotic polynomial as described in claim 7, characterized in that: The specific implementation method of step eight is as follows: Based on the dynamically adjusted damage determination constraints constructed in step seven Determine the damage requirements for this group of targets; the objective of aiming point allocation is to maximize the damage effect while satisfying the damage requirements. Aiming point optimization is performed in groups, therefore the optimization model is as follows: Among them, D i The damage effect under the current aiming point scheme; x L x U y L and y U This refers to the range of the aiming point. The final hierarchical firepower planning result is obtained by solving equation (21).

9. A hierarchical firepower planning method for a fusion diffusion model and a chaotic polynomial as described in claims 2, 3, 4, 5, 6, 7, or 8, characterized in that: It also includes step nine, which, based on the final hierarchical fire planning results obtained in step eight, accurately identifies the importance and threat level of different targets, accurately assesses the actual damage caused by the aircraft to the targets by considering CEP and TEP, matches the mission planning results with the actual results, and ensures that the aircraft aiming point can be obtained quickly in the real mission environment by optimizing the fire planning model.

Images